A Purchase Risk Assessment Method Based on User Behavior Habits

Abstract

This invention relates to the field of purchase risk assessment technology and discloses a purchase risk assessment method based on user behavior habits. The method includes: obtaining the user's registration time and dividing the user's behavior into two time periods based on the registration time: the period within the last seven days of registration as the first time period, and all time periods within the registration time excluding the last seven days as the second time period; firstly, comparing the average number of purchases, purchase frequency, and purchase amount in the first and second time periods to determine the user's risk level; secondly, determining the user's risk level based on the time the user spends browsing similar products, payment methods, delivery methods, and review methods; finally, determining the user's overall risk level based on the risk levels of each aspect, and accordingly determining whether the user poses a purchase risk. This solution helps merchants effectively and accurately identify user purchase risks and helps reduce transaction risks for both merchants and users.

Description

Technical Field

[0001] This invention relates to the field of purchase risk assessment technology, specifically a purchase risk assessment method based on user behavior habits. Background Technology

[0002] With the continuous advancement of technology and changes in the social environment, the development of online shopping has been driven by a combination of economic, technological, and socio-cultural factors. As a result, more and more consumers are choosing to purchase goods online, and online shopping will continue to be an important way for consumers to shop, constantly evolving to adapt to new market demands as online shopping rapidly develops.

[0003] However, the rapid development of online shopping has also brought a series of challenges, especially in terms of merchants assessing the purchase risks of users based on their purchasing behavior. In order to protect the interests of both users and merchants, effective risk assessment methods need to be adopted.

[0004] Traditional risk assessment methods often focus on a single indicator or a limited number of indicators, such as purchase amount, purchase frequency, and average purchase amount, while ignoring the complexity and diversity of users' behavioral habits when making purchases, as well as other factors that may affect risk assessment, such as changes in users' payment methods and delivery addresses.

[0005] In traditional risk assessment processes, shopping platforms assume that different types of risks have the same priority, which may lead to the erroneous labeling of users who do not pose a purchase risk as high-risk users, thus hindering transactions between merchants and users. Summary of the Invention

[0006] This invention provides a purchase risk assessment method based on user behavior habits, which helps to solve the problems mentioned in the background art.

[0007] This invention provides the following technical solution: a method for assessing purchase risk based on user behavior habits.

[0008] Optionally, a purchase risk assessment method based on user behavior habits is characterized by including:

[0009] Get the user's registration time D on the shopping platform, and get the number of purchases X1 in the first time period and the number of purchases X2 in the second time period;

[0010] The first time period is the period within the last seven days of the registration time, and the second time period is all time periods within the registration time excluding the last seven days.

[0011] Obtain the user's purchase frequency, compare the user's purchase frequency in the first time period with the user's purchase frequency in the second time period, and determine the user's risk level;

[0012] Obtain the user's total purchase amount Y1 for the first time period and the total purchase amount Y2 for the second time period;

[0013] Calculate the average purchase amount, compare the average purchase amount of the user in the first time period with the average purchase amount of the user in the second time period, and determine the user's risk level;

[0014] The system obtains the time a user spends browsing similar products within a single purchase period in the first time period, and the time of that single purchase period. Based on the time a user spends browsing similar products within that single purchase period in the first time period, the system determines the user's risk level.

[0015] By obtaining users' payment methods, delivery methods, product review times, and product ratings, we can determine whether users pose a purchase risk.

[0016] The system identifies the merchants who transacted with a user within a specific time period. Based on the purchasing behavior of all other users who transacted with these merchants within the same time period, the system determines the user's risk level.

[0017] The risk level P of the user is used to determine whether the user faces any purchase risk.

[0018] Optionally, obtaining the user's purchase frequency involves comparing the user's purchase frequency in a first time period with the user's purchase frequency in a second time period to determine the user's risk level. Specifically:

[0019] The purchase frequency is specifically calculated as the ratio of the number of purchases to the time spent completing the purchase.

[0020] To obtain the user's purchase frequency A1 in the first time period, specifically the ratio of the number of purchases X1 in the first time period to the time 7 spent completing the purchase in the first time period, A1 = X1 ÷ 7.

[0021] To obtain the user's purchase frequency A2 in the second time period, specifically the ratio of the number of purchases X2 in the second time period to the time spent completing the purchase in the second time period (D-7), we can calculate A2 = X2 ÷ (D-7).

[0022] Compare the values ​​of A1 and A2. If the value of A1 is greater than the value of A2, the user's risk level is recorded as P1. If the value of A1 is less than or equal to the value of A2, the user is not marked.

[0023] Optionally, the calculation of the average purchase amount involves comparing the average purchase amount of the user in the first time period with the average purchase amount of the user in the second time period to determine the user's risk level. Specifically:

[0024] The average purchase amount is specifically calculated as the ratio of the total purchase amount to the number of purchases;

[0025] Get the average purchase amount B1 of the user in the first time period, which is the ratio of the total purchase amount Y1 of the user in the first time period to the number of purchases X1 in the first time period. Then B1 = Y1 ÷ X1.

[0026] To obtain the average purchase amount B2 of the user in the second time period, specifically the ratio of the total purchase amount Y2 of the user in the second time period to the number of purchases X2 of the user in the second time period, we have B2 = Y2 ÷ X2.

[0027] Compare the values ​​of B1 and B2. If the value of B1 is greater than the value of B2, the user's risk level is recorded as P2. If the value of B1 is less than or equal to the value of B2, the user is not marked.

[0028] Optionally, the step of obtaining the time spent by the user browsing similar products within a single purchase cycle of the first time period, obtaining the time of the user's single purchase cycle within the first time period, and determining the user's risk level based on the time spent by the user browsing similar products within the single purchase cycle of the first time period and the time of the user's single purchase cycle within the first time period, specifically involves:

[0029] Get the number of purchases made by the user in the first time period, and record the number as N. Then, according to the time order, record the N purchases as the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle.

[0030] The purchase cycle is specifically the time elapsed from the start of searching for the product to the final completion of the purchase.

[0031] S1. Obtain the number M of similar products viewed by the user within a single purchase cycle, and the time spent viewing each product when viewing M similar products;

[0032] S2. Let T1, T2, T3, ..., Tm be the time spent browsing the first, second, third, ..., Mth items respectively. M ;

[0033] S3, Calculate T1, T2, T3...T M Find the difference between each pair of pairs, and obtain the absolute value of all the differences. The number of absolute values ​​is denoted as C, where C = [M(M-1) / 2].

[0034] S4. Set a value T, and count the number of values ​​less than or equal to T among the C absolute values ​​obtained. The number is denoted as R.

[0035] S5. Calculate the ratio of R to C, denoted as E. The specific calculation is E = R ÷ C = R ÷ [M(M-1) / 2];

[0036] Let E1, E2, E3...E be the ratio of the number of absolute values ​​greater than C in the first purchase period, the second purchase period, the third purchase period, ... the Nth purchase period to the total number of absolute values ​​obtained. N ;

[0037] Set a value F, and count the values ​​of E1, E2, E3...E N The number of values ​​greater than F is denoted as n. Calculate the ratio of n to N.

[0038] Set a value Z1. If the ratio of n to N is greater than Z1, the user's risk level is recorded as P3.

[0039] If the ratio of n to N is less than or equal to Z1, then the user will not be marked.

[0040] Optionally, obtaining the user's payment method, delivery method, product review time, and product review rating to determine whether the user poses a purchase risk specifically involves:

[0041] Obtain the payment methods used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number of different payment methods G, and calculate the ratio of G to N;

[0042] Obtain the shipping addresses used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number H of different shipping addresses, and calculate the ratio of H to N;

[0043] If the ratio of G to N is greater than a value Z2, and the ratio of H to N is greater than a value Z3, the user's risk level is recorded as P4.

[0044] If the ratio of G to N is less than or equal to a value Z2, or the ratio of H to N is less than or equal to a value Z3, then the user will not be marked.

[0045] Optionally, the step of obtaining the user's payment method, delivery method, evaluation time of the product, and evaluation rating of the product to determine whether the user has a purchase risk also includes:

[0046] Get user reviews of products during the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period;

[0047] The rating system refers to the process whereby, after a user completes a purchase, they rate the product based on their satisfaction with it. The ratings are then divided into five levels from low to high based on the level of satisfaction: one-star, two-star, three-star, four-star, and five-star.

[0048] Count the number U of five-star ratings out of the N ratings obtained, and calculate the ratio of U to N;

[0049] Obtain the evaluation time values ​​for the first purchase cycle, the second purchase cycle, the third purchase cycle, ..., the Nth purchase cycle, and denote them as K1, K2, K3...K respectively. N ;

[0050] Let K1, K2, K3...K N Take the difference between each pair of pairs and obtain the absolute value of all the differences. The number of absolute values ​​is denoted as L, where L = [N(N-1) / 2].

[0051] Set a value Z4, count the number of absolute values ​​obtained that are less than or equal to the value Z4, and denote the number as V. Calculate the ratio of V to N.

[0052] If the ratio of U to N is greater than a value Z5, and the ratio of V to N is greater than a value Z6, the user's risk level is recorded as P5.

[0053] If the ratio of U to N is less than or equal to a value Z5, or the ratio of V to N is less than or equal to a value Z6, then the user will not be marked.

[0054] Optionally, the step of obtaining the merchants who transacted with the user within the first time period, and determining the user's risk level based on the purchasing behavior of all other users who transacted with these merchants within the first time period, specifically involves:

[0055] Get the number of merchants W1 that transacted with the user in the first time period;

[0056] If, within the first time period, a risky user transacts with one of the W1 merchants, then this merchant is marked as a risky merchant, and the number of risky merchants among the W1 merchants is denoted as W2. The ratio of W2 to W1 is then calculated.

[0057] The term "risk user" specifically refers to a user whose risk level is any one or more of P1, P2, P3, P4, or P5.

[0058] If a value Z7 is set, and the ratio of W2 to W1 is greater than the value Z7, the user's risk level is recorded as P6.

[0059] If the ratio of W2 to W1 is less than or equal to Z7, the user will not be marked.

[0060] Optionally, the step of determining whether a user faces purchase risk based on their overall risk level P specifically involves:

[0061] The total risk level P is specifically the sum of the user's risk level values, specifically: if the user's risk is P1, P3, P4 and P5, then the user's total risk level P = P1 + P3 + P4 + P5.

[0062] Where P1+P2+P3+P4+P5+P6=1, and P3>P5>P4>P6>P1>P2;

[0063] If a user's overall risk level P is greater than a value Z8, the user is considered to have a purchase risk; if a user's overall risk level P is less than or equal to a value Z8, the user is considered to have no purchase risk.

[0064] The present invention has the following beneficial effects:

[0065] 1. This purchase risk assessment method based on user behavior habits determines the user's risk level by comparing the user's purchase frequency and average purchase amount in the first and second time periods. By comparing the user's purchase behavior in the two time periods, abnormal user behavior can be identified more accurately, thereby improving the accuracy of risk assessment.

[0066] 2. This purchase risk assessment method based on user behavior habits obtains the time difference of a user browsing similar products within a single purchase cycle. By carefully analyzing the time difference of users browsing similar products, it can more accurately assess the user's purchase intention and potential risks, and determine whether there are any potential abnormal situations. Combined with the time spent in each of the N purchase cycles, it can effectively and accurately determine whether there are any purchase risks during the purchase process, which helps to reduce the transaction risks for merchants.

[0067] 3. This purchase risk assessment method based on user behavior habits combines the user's payment method with the user's delivery method when determining whether the user has a purchase risk; and combines the user's evaluation time with the evaluation time when determining whether the user has a purchase risk based on the user's evaluation rating of the product. This reduces the misjudgment of normal users and improves the accuracy of risk assessment.

[0068] 4. This purchase risk assessment method based on user behavior habits determines whether a user has a purchase risk from multiple aspects, and different aspects of purchase risk have different priorities. This can reduce overreaction to low-risk user purchase behaviors and also reduce erroneous judgments of low-risk user purchase behaviors, thereby improving identification efficiency. Attached Figure Description

[0069] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] Example 1, refer to Figure 1 A method for assessing purchase risk based on user behavior habits, characterized by: including:

[0072] Get the user's registration time D on the shopping platform, where the registration time D is in days. Get the number of purchases X1 in the first time period and the number of purchases X2 in the second time period.

[0073] The first time period is the period within the last seven days of the registration time, and the second time period is all time periods within the registration time excluding the last seven days.

[0074] Obtain the user's purchase frequency, compare the user's purchase frequency in the first time period with the user's purchase frequency in the second time period, and determine the user's risk level;

[0075] Obtain the user's total purchase amount Y1 for the first time period and the total purchase amount Y2 for the second time period;

[0076] Calculate the average purchase amount and compare the average purchase amount of the user in the first time period with the average purchase amount of the user in the second time period to determine the user's risk level. By comparing the user's purchase frequency and average purchase amount in the two time periods, abnormal user behavior can be identified more accurately, thereby improving the accuracy of risk assessment.

[0077] The system obtains the time a user spends browsing similar products within a single purchase period in the first time period, and the time of that single purchase period. Based on the time a user spends browsing similar products within that single purchase period in the first time period, the system determines the user's risk level.

[0078] By obtaining users' payment methods, delivery methods, product review times, and product ratings, we can determine whether users pose a purchase risk.

[0079] The system identifies the merchants who transacted with a user within a specific time period. Based on the purchasing behavior of all other users who transacted with these merchants within the same time period, the system determines the user's risk level.

[0080] The risk level P of the user is used to determine whether the user faces any purchase risk.

[0081] The process of obtaining a user's purchase frequency involves comparing the user's purchase frequency in a first time period with the user's purchase frequency in a second time period to determine the user's risk level. Specifically:

[0082] The purchase frequency is specifically calculated as the ratio of the number of purchases to the time spent completing the purchase.

[0083] Get the user's purchase frequency A1 in the first time period, which is the ratio of the number of purchases X1 in the first time period to the time 7 spent completing the purchase in the first time period, where 7 is the number of days in the first time period. Then A1 = X1 ÷ 7.

[0084] To obtain the user's purchase frequency A2 in the second time period, specifically, it is the ratio of the number of purchases X2 in the second time period to the time (D-7) spent completing the purchase in the second time period, where (D-7) is the number of days in the second time period. Therefore, A2 = X2 ÷ (D-7).

[0085] Compare the values ​​of A1 and A2. If the value of A1 is greater than the value of A2, it means that the user's purchase frequency in the first time period is higher than that in the second time period. This indicates that the user's purchase frequency in the first time period is abnormal, and the user's risk level is recorded as P1. If the value of A1 is less than or equal to the value of A2, it means that the user's purchase frequency in the first time period is not higher than that in the second time period. This indicates that the user's purchase frequency in the first time period is not abnormal, and the user is not marked.

[0086] The calculation of the average purchase amount involves comparing the user's average purchase amount in the first time period with the user's average purchase amount in the second time period to determine the user's risk level. Specifically:

[0087] The average purchase amount is specifically calculated as the ratio of the total purchase amount to the number of purchases;

[0088] Get the average purchase amount B1 of the user in the first time period, which is the ratio of the total purchase amount Y1 of the user in the first time period to the number of purchases X1 in the first time period. Then B1 = Y1 ÷ X1.

[0089] To obtain the average purchase amount B2 of the user in the second time period, specifically the ratio of the total purchase amount Y2 of the user in the second time period to the number of purchases X2 of the user in the second time period, we have B2 = Y2 ÷ X2.

[0090] Compare the values ​​of B1 and B2. If the value of B1 is greater than the value of B2, it means that the average purchase amount of the user in the first time period is higher than the purchase frequency in the second time period. This indicates that the average purchase amount of the user in the first time period is abnormal, and the user's risk level is recorded as P2. If the value of B1 is less than or equal to the value of B2, it means that the average purchase amount of the user in the first time period is not higher than the purchase frequency in the second time period. This indicates that the average purchase amount of the user in the first time period is not abnormal, and the user is not marked.

[0091] The process involves obtaining the time spent browsing similar products within a single purchase cycle of the first time period, obtaining the time of that single purchase cycle, and determining the user's risk level based on the time spent browsing similar products within that single purchase cycle and the time of that single purchase cycle. Specifically:

[0092] Get the number of purchases made by the user in the first time period, and record the number as N. Then, according to the time order, record the N purchases as the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle.

[0093] The purchase cycle is specifically the time elapsed from the start of searching for the product to the final completion of the purchase.

[0094] S1. Obtain the number M of similar products viewed by the user within a single purchase cycle, and the time spent viewing each product when viewing M similar products;

[0095] S2. Let T1, T2, T3, ..., Tm be the time spent browsing the first, second, third, ..., Mth items respectively. M ;

[0096] S3, Calculate T1, T2, T3...T M The difference between each pair is used to obtain the degree of closeness between different times and to obtain the absolute value of all differences. The number of absolute values ​​is denoted as C, where C = [M(M-1) / 2].

[0097] S4. Set a value T. The function of value T is to judge the closeness between two individual purchase cycle times. When T is too large, it is impossible to accurately obtain the closeness between two individual purchase cycle times. The smaller T is, the better it reflects the closeness between two individual purchase cycle times. When T = 0, the two individual purchase cycle times are completely equal. Count the number of values ​​less than or equal to T among the C absolute values ​​obtained. The number is denoted as R.

[0098] S5. Calculate the ratio of R to C, denoted as E. The ratio E reflects the proportion of the differences that are close in value to the total number of differences. The specific calculation is E = R ÷ C = R ÷ [M(M-1) / 2];

[0099] Let E1, E2, E3...E be the ratio of the number of absolute values ​​greater than C in the first purchase period, the second purchase period, the third purchase period, ... the Nth purchase period to the total number of absolute values ​​obtained. N ;

[0100] Set a value F, and count the values ​​of E1, E2, E3...E N The number of values ​​greater than F, if E1, E2, E3...E N If there is a ratio among N ratios that is greater than F, it means that the ratios with similar values ​​account for a relatively high proportion of all differences. Let the number of such ratios be n. Calculate the ratio of n to N.

[0101] Set a value Z1. The purpose of Z1 is to determine whether there is any abnormality in the time spent by the user browsing similar products or a single purchase cycle. If the ratio of n to N is greater than Z1, that is, there is an abnormality in the time spent by the user browsing similar products or a single purchase cycle, the user's risk level is recorded as P3.

[0102] If the ratio of n to N is less than or equal to Z1, meaning that there is no abnormality in the time spent by the user browsing similar products or in a single purchase cycle, then the user will not be flagged.

[0103] The process of obtaining the user's payment method, delivery method, product review time, and product review rating to determine whether the user poses a purchase risk specifically involves:

[0104] Obtain the payment methods used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number of different payment methods G, and calculate the ratio of G to N;

[0105] Obtain the shipping addresses used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number H of different shipping addresses, and calculate the ratio of H to N;

[0106] If the ratio of G to N is greater than a value Z2, it means that the user's payment method is abnormal. The value Z2 is used to determine whether the user's payment method is abnormal based on the relationship between the ratio of G to N and the value Z2. If the ratio of H to N is greater than a value Z3, it means that the user's payment method is abnormal. The value Z3 is used to determine whether the user's payment method is abnormal based on the relationship between the ratio of H to N and the value Z3. The user's risk level is recorded as P4.

[0107] If the ratio of G to N is less than or equal to a value Z2, or the ratio of H to N is less than or equal to a value Z3, meaning that the user's payment method or the user's receiving method is not abnormal, then the user will not be marked.

[0108] The method of obtaining the user's payment method, delivery method, product review time, and product review rating to determine whether the user poses a purchase risk also includes:

[0109] Get user reviews of products during the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period;

[0110] The rating system refers to the process whereby, after a user completes a purchase, they rate the product based on their satisfaction with it. The ratings are then divided into five levels from low to high based on the level of satisfaction: one-star, two-star, three-star, four-star, and five-star.

[0111] Count the number U of five-star ratings out of the N ratings obtained, and calculate the ratio of U to N;

[0112] Obtain the evaluation time values ​​for the first purchase cycle, the second purchase cycle, the third purchase cycle, ..., the Nth purchase cycle, and denote them as K1, K2, K3...K respectively. N ;

[0113] Let K1, K2, K3...K N Take the difference between each pair of pairs and obtain the absolute value of all the differences. The number of absolute values ​​is denoted as L, where L = [N(N-1) / 2].

[0114] Set a value Z4. The function of Z4 is to judge the closeness between two evaluation times. When Z4 is too large, it is impossible to accurately obtain the closeness between the two evaluation times. The smaller Z4 is, the better it reflects the closeness between the two evaluation times. When Z4 = 0, the two evaluation times are completely equal. Count the number of values ​​less than or equal to a value Z4 among the L absolute values ​​obtained. Record the number as V. Calculate the ratio of V to N.

[0115] If the ratio of U to N is greater than a value Z5, it means that the user's rating level is abnormal. The value Z5 is used to determine whether the user's rating level is abnormal based on the relationship between the ratio of U to N and the value Z5. If the ratio of V to N is greater than a value Z6, it means that the user's evaluation time is abnormal. The value Z6 is used to determine whether the user's evaluation time is abnormal based on the relationship between the ratio of V to N and the value Z6. The user's risk level is recorded as P5.

[0116] If the ratio of U to N is less than or equal to a value Z5, or the ratio of V to N is less than or equal to a value Z6, meaning that the user's rating level or rating time is not abnormal, then the user will not be marked.

[0117] The process of obtaining merchants who transacted with a user within a first time period, and determining the user's risk level based on the purchasing behavior of all other users who transacted with these merchants within the first time period, specifically involves:

[0118] Get the number of merchants W1 that transacted with the user in the first time period;

[0119] If, within the first time period, a risky user transacts with one of the W1 merchants, then this merchant is marked as a risky merchant, and the number of risky merchants among the W1 merchants is denoted as W2. The ratio of W2 to W1 is then calculated.

[0120] The term "risk user" specifically refers to a user whose risk level is any one or more of P1, P2, P3, P4, or P5.

[0121] A value Z7 is set. The purpose of Z7 is to determine whether a user is abnormal based on the number of risky merchants among those who transacted with the user in the first time period. If the value Z7 is too large, it means that the number of risky merchants among those who transacted with the user in the first time period is high enough to be judged as having purchase risk, which may cause some users who actually have high risk to not be identified in time, thus reducing the accuracy of risk identification. If the value Z7 is too small, even if the number of risky merchants among those who transacted with the user in the first time period is low, the user may still be marked as having purchase risk, which will lead to an increase in the error rate of purchase risk identification and affect the shopping experience of normal users. If the ratio of W2 to W1 is greater than the value Z7, it means that the number of risky merchants among those who transacted with the user in the first time period is large, then this user is abnormal, and the user's risk level is recorded as P6.

[0122] If the ratio of W2 to W1 is less than or equal to Z7, it means that there are relatively few risky merchants among the merchants who transacted with the user in the first time period. Therefore, this user is not abnormal and will not be marked.

[0123] The method of determining whether a user faces purchase risk based on their overall risk level P is as follows:

[0124] The total risk level P is specifically the sum of the user's risk level values, specifically: if the user's risk is P1, P3, P4 and P5, then the user's total risk level P = P1 + P3 + P4 + P5.

[0125] Where P1+P2+P3+P4+P5+P6=1, and P3>P5>P4>P6>P1>P2, that is, different risk levels have different priorities, which can reduce overreaction to users' low-risk purchase behavior, reduce misjudgment of users' low-risk behavior, and improve identification efficiency.

[0126] If a user's overall risk level P is greater than a value Z8, then a large Z8 indicates a sufficiently high risk level to be considered a potential purchase risk, which may lead to some users with actual high risk not being identified in time, thus reducing the accuracy of risk identification. Conversely, a small Z8 may indicate that even users with low risk levels are marked as having a purchase risk, leading to an increased error rate in risk identification and affecting the shopping experience of normal users. A user is considered to have a purchase risk if they are deemed to have one, and if their overall risk level P is less than or equal to a value Z8, they are considered to have no purchase risk.

[0127] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0128] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing purchase risk based on user behavior habits, characterized in that: include: Get the user's registration time D on the shopping platform, and get the number of purchases X1 in the first time period and the number of purchases X2 in the second time period; The first time period is the period within the last seven days of the registration time, and the second time period is all time periods within the registration time excluding the last seven days. Obtain the user's purchase frequency, compare the user's purchase frequency in the first time period with the user's purchase frequency in the second time period, and determine the user's risk level; Obtain the user's total purchase amount Y1 for the first time period and the total purchase amount Y2 for the second time period; Calculate the average purchase amount, compare the average purchase amount of the user in the first time period with the average purchase amount of the user in the second time period, and determine the user's risk level; The system obtains the time a user spends browsing similar products within a single purchase period in the first time period, and the time of that single purchase period. Based on the time a user spends browsing similar products within that single purchase period in the first time period, the system determines the user's risk level. Get the number of purchases made by the user in the first time period, and record the number as N. Then, according to the time order, record the N purchases as the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle. The purchase cycle is specifically the time elapsed from the start of searching for the product to the final completion of the purchase. S1. Obtain the number M of similar products viewed by the user within a single purchase cycle, and the time spent viewing each product when viewing M similar products; S2. Let T1, T2, T3, ..., Tm be the time spent browsing the first, second, third, ..., Mth items respectively. M ; S3, Calculate T1, T2, T3...T M Find the difference between each pair of pairs of pairs, and obtain the absolute value of all the differences. The number of absolute values ​​is denoted as C, where C = [M(M-1) / 2]. S4. Set a value T, and count the number of values ​​less than or equal to T among the C absolute values ​​obtained. The number is denoted as R. S5. Calculate the ratio of R to C, denoted as E. The specific calculation is E = R ÷ C = R ÷ [M(M-1) / 2]; Let E1, E2, E3...E be the ratio of the number of absolute values ​​greater than C in the first purchase period, the second purchase period, the third purchase period, ... the Nth purchase period to the total number of absolute values ​​obtained. N ; Set a value F, and count the values ​​of E1, E2, E3...E N The number of values ​​greater than F is denoted as n. Calculate the ratio of n to N. Set a value Z1. If the ratio of n to N is greater than Z1, the user's risk level is recorded as P3. If the ratio of n to N is less than or equal to Z1, then the user will not be marked. By obtaining users' payment methods, delivery methods, product review times, and product ratings, we can determine whether users pose a purchase risk. The system identifies the merchants who transacted with a user within a specific time period. Based on the purchasing behavior of all other users who transacted with these merchants within the same time period, the system determines the user's risk level. The risk level P of the user is used to determine whether the user faces any purchase risk.

2. The purchase risk assessment method based on user behavior habits according to claim 1, characterized in that: The process of obtaining a user's purchase frequency involves comparing the user's purchase frequency in a first time period with the user's purchase frequency in a second time period to determine the user's risk level. Specifically: The purchase frequency is specifically calculated as the ratio of the number of purchases to the time spent completing the purchase. To obtain the user's purchase frequency A1 in the first time period, specifically the ratio of the number of purchases X1 in the first time period to the time 7 spent completing the purchase in the first time period, A1 = X1 ÷ 7. To obtain the user's purchase frequency A2 in the second time period, specifically the ratio of the number of purchases X2 in the second time period to the time spent completing the purchase (D-7) in the second time period, we can calculate A2 = X2 ÷ (D-7). Compare the values ​​of A1 and A2. If the value of A1 is greater than the value of A2, the user's risk level is recorded as P1. If the value of A1 is less than or equal to the value of A2, the user is not marked.

3. The purchase risk assessment method based on user behavior habits according to claim 1, characterized in that: The calculation of the average purchase amount involves comparing the user's average purchase amount in the first time period with the user's average purchase amount in the second time period to determine the user's risk level. Specifically: The average purchase amount is specifically calculated as the ratio of the total purchase amount to the number of purchases; Get the average purchase amount B1 of the user in the first time period. Specifically, it is the ratio of the total purchase amount Y1 of the user in the first time period to the number of purchases X1 in the first time period. Then B1 = Y1 ÷ X1. To obtain the average purchase amount B2 of the user in the second time period, specifically the ratio of the total purchase amount Y2 of the user in the second time period to the number of purchases X2 of the user in the second time period, we have B2 = Y2 ÷ X2. Compare the values ​​of B1 and B2. If the value of B1 is greater than the value of B2, the user's risk level is recorded as P2. If the value of B1 is less than or equal to the value of B2, the user is not marked.

4. The purchase risk assessment method based on user behavior habits according to claim 1, characterized in that: The process of obtaining the user's payment method, delivery method, product review time, and product review rating to determine whether the user poses a purchase risk specifically involves: Obtain the payment methods used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number of different payment methods G, and calculate the ratio of G to N; Obtain the shipping addresses used by the user in the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period, count the number H of different shipping addresses, and calculate the ratio of H to N; If the ratio of G to N is greater than a value Z2, and the ratio of H to N is greater than a value Z3, the user's risk level is recorded as P4. If the ratio of G to N is less than or equal to a value Z2, or the ratio of H to N is less than or equal to a value Z3, then the user will not be marked.

5. The purchase risk assessment method based on user behavior habits according to claim 1, characterized in that: The method of obtaining the user's payment method, delivery method, product review time, and product review rating to determine whether the user poses a purchase risk also includes: Get user reviews of products during the first purchase cycle, the second purchase cycle, the third purchase cycle, ... the Nth purchase cycle within the first time period; The rating system refers to the process whereby, after a user completes a purchase, they rate the product based on their satisfaction with it. The ratings are then divided into five levels from low to high based on the level of satisfaction: one-star, two-star, three-star, four-star, and five-star. Count the number U of five-star ratings out of the N ratings obtained, and calculate the ratio of U to N; Obtain the evaluation time values ​​for the first purchase cycle, the second purchase cycle, the third purchase cycle, ..., the Nth purchase cycle, and denote them as K1, K2, K3...K respectively. N ; Let K1, K2, K3...K N Take the difference between each pair of pairs and obtain the absolute value of all the differences. The number of absolute values ​​is denoted as L, where L = [N(N-1) / 2]. Set a value Z4, count the number of absolute values ​​less than or equal to Z4 among the L obtained values, denoted as V, and calculate the ratio of V to N. If the ratio of U to N is greater than a value Z5, and the ratio of V to N is greater than a value Z6, the user's risk level is recorded as P5. If the ratio of U to N is less than or equal to a value Z5, or the ratio of V to N is less than or equal to a value Z6, then the user will not be marked.

6. The method for assessing purchase risk based on user behavior habits according to claim 1, characterized in that: The process of obtaining merchants who transacted with a user within a first time period, and determining the user's risk level based on the purchasing behavior of all other users who transacted with these merchants within the first time period, specifically involves: Get the number of merchants W1 that transacted with the user in the first time period; If, within the first time period, a risky user transacts with one of the W1 merchants, then this merchant is marked as a risky merchant, and the number of risky merchants among the W1 merchants is denoted as W2. The ratio of W2 to W1 is then calculated. The term "risk user" specifically refers to a user whose risk level is any one or more of P1, P2, P3, P4, or P5. If a value Z7 is set, and the ratio of W2 to W1 is greater than the value Z7, the user's risk level is recorded as P6. If the ratio of W2 to W1 is less than or equal to Z7, the user will not be marked.

7. The purchase risk assessment method based on user behavior habits according to claim 1, characterized in that: The method of determining whether a user faces purchase risk based on their overall risk level P is as follows: The total risk level P is specifically the sum of the user's risk level values, specifically: if the user's risk is P1, P3, P4 and P5, then the user's total risk level P = P1 + P3 + P4 + P5. Where P1+P2+P3+P4+P5+P6=1, and P3>P5>P4>P6>P1>P2; If a user's overall risk level P is greater than a value Z8, the user is considered to have a purchase risk; if a user's overall risk level P is less than or equal to a value Z8, the user is considered to have no purchase risk.

Images